One of the challenges of wireless communications is that there is a finite amount of bandwidth available over the air, but an ever increasing population of mobile devices trying to access it. Through scheduling requests, devices are able to share radio resources and, when they require access to these resources, can be granted access to use them. Many wireless communication systems implement some form of scheduling requests, however it is emerging that these current solutions are unable to cope with the changing behaviour and requirements of modern mobile devices.
Long Term Evolution (LTE), a standard for high-speed wireless communication, comprises an Evolved UMTS Terrestrial Radio Access Network (E-UTRAN) coupled to an Evolved Packet Core (EPC) Network. In FIG. 1 the E-UTRAN 103 comprises one node type eNB 102 whilst the EPC 104 comprises 3 node types. Whilst the term eNB is used to denote the access node used for the Uu interface (interface that links the UE to the E-UTRAN), there are a variety of access nodes capable of operating on the Uu interface, such as relays, home eNode Bs, etc. Therefore, for the purposes of describing the embodiments herein, the term eNB may be used to refer to either the eNB, or to other access nodes performing equivalent or similar operations on the Uu interface. The Serving Gateway (SGW) 105 routes user-plane data within the core network, the Mobility Management Endpoint (MME) 106 handles mobility and connection control between the UE and the core network, and the Packet Gateway (PGW) 107 ingress/egress node routes data between the core network and external networks. FIG. 1 also shows the network interfaces between nodes
The LTE system has 3 main uplink physical channel types: Physical Random Access Channel (PRACH), Physical Uplink Control Channel (PUCCH) and Physical Uplink Shared Channel (PUSCH),
The Physical Random Access Channel (PRACH) is a contention-based channel where transmissions from multiple users need only be very loosely synchronised (time of arrival differences may be of the order of 0.1 s or more). Devices using PRACH normally communicate on an ad-hoc basis.
PUCCH and PUSCH orthogonal resources, on the other hand, require tighter synchronicity of devices (time of arrival difference of order of 5 μs or less within the cyclic prefix duration of a Single Carrier Frequency Division Multiple Access ‘SC-FDMA’ symbol) and therefore allow for multiple devices to use shared resources within the uplink system bandwidth.
SC-FDMA
SC-FDMA is the modulation scheme used in the uplink of LTE. The scheme has some similarities to Orthogonal Frequency Division Multiple Access (OFDMA) but also some key differences. In both OFDMA and SC-FDMA, the physical resource is sub-divided into a time-frequency grid of Resource Elements (REs), each RE consisting of one unit of frequency (one sub-carrier) and one unit of time (an OFDMA or SC-FDMA symbol duration). The OFDMA or SC-FDMA symbols are both pre-pended with a cyclic prefix (CP) (an end portion of the symbol is copied and inserted at the start of the transmitted symbol).
In both SC-FDMA and OFDMA, information bits to be transmitted may undergo steps of encoding (to form encoded bits) and the encoded bits are usually then mapped to data modulation symbols (for example, using QPSK, 16-QAM or 64-QAM modulation schemes). A key difference between OFDMA and SC-FDMA concerns how these data modulation symbols are mapped onto the time-frequency resources.
In the case of OFDMA, there is usually a one-to-one mapping of data modulation symbols onto sub-carriers. That is, N modulation symbols to be transmitted within one OFDMA symbol duration are each mapped to a corresponding one of a set of N sub-carriers.
Conversely, in the case of SC-FDMA, there is usually a one-to-many mapping of data modulation symbols to sub-carriers. That is, one data modulation symbol is mapped via a spreading operation to a set of (usually complex) numerical values, and each of these numerical values is then used to modulate a particular one of the N sub-carriers. Furthermore, other data modulation symbols may be mapped via a similar (and usually co-ordinated) spreading operation to the same set of (usually consecutive) N sub-carriers.
The spreading operations are usually co-ordinated in order to ensure that the multiple data modulation symbols remain orthogonal to one another. Therefore, the data modulation symbols may be transmitted from multiple devices and received with simple linear operations with low inter-device interference. This behaviour can increase spectral efficiency by allowing multiple devices with small amounts of information to transmit to share a set of subcarriers.
When multiple modulation symbols in one device are spread over the same set of subcarriers, each modulation symbol creates an additional set of N numerical values, and these are linearly combined (across the combinations from each of the contributing modulation symbols) to form the eventual numerical values that are used to modulate each sub-carrier. This linear combination of spread symbols can be viewed as a transform operation. The transform operation is usually designed to ensure that the resultant transmitted signal has a lower Peak-to-Average Power Ratio (PAPR) than its OFDMA counterpart. The transform operation may comprise a Discrete Fourier Transform (DFT) as is often used in the case of uplink transmissions for LTE, but other transforms are also possible that preserve a low PAPR of the transmitted signal. Similarly, when only one modulation symbol (e.g. either a data modulation symbol or a reference symbol) is transmitted over a set of sub-carriers, the spreading sequence is also designed to minimize its PAPR. In this case, spreading sequences with constant (or near constant) amplitude and zero (or low) autocorrelation (“CAZAC”) properties are often used such as when constructing uplink reference signals or uplink physical control channel transmissions in LTE.
Thus, a data modulation symbol is transmitted over a set of N sub-carriers for SC-FDMA, whereas one data modulation symbol is transmitted over one sub-carrier in the case of OFDMA.
PUCCH and PUSCH
FIG. 2 shows an example structure of a sub-frame 210 spanning 1 ms in the time dimension and an uplink system bandwidth in the frequency dimension. The sub-frame comprises two slots of duration 0.5 ms each. The sub-frame 210 can be pictured as a collection of discrete blocks 220, each comprising 12 sub-carriers of 15 kHz each in the frequency domain and a single SC-FDMA symbol 260 in time. In this example, each resource block 250 (of PUSCH or PUCCH resource) comprises 12×14 resource elements (RE) within a sub-frame, where a RE is one sub-carrier unit of frequency and one SC-FDMA symbol 260 of time resource. The number of SC-FDMA symbols 260 per sub-frame may vary depending upon system configuration, thereby also affecting the number of REs per sub-frame. The system configuration may be a function of a cyclic prefix (CP) duration of the SC-FDMA symbols. The PUSCH resource 240 is located in the central frequency region with the PUCCH control regions 230 at the edges above and below. Within a resource block 250, certain SC-FDMA symbols may be used for reference signal (RS) purposes. Reference signals are signals known to the receiver and which may be used for estimation of the radio channel in order to improve demodulation and detection performance. In the example of FIG. 2, the 4th SC-FDMA symbol of each slot is used for PUSCH RS. The symbol locations of RS for the PUCCH region may vary as a function of a PUCCH signal format. For example, for a PUCCH format 1 signal, the RS may be located on the 3rd, 4th and 5th SC-FDMA symbols of each slot, whereas for a PUCCH format 2 or PUCCH format 3 signal, the RS may be located on the 2nd and 6th SC-FDMA symbols of each slot.
Synchronised mobile devices know the time and frequency locations of these resources, hence the PUSCH resources 240 can be dynamically shared under the control of a scheduler (allocated using uplink grants within Downlink Control Information (DCI) messages sent on the Physical Downlink Control Channel (PDCCH). Generally it is within this PUSCH 240 resource that uplink user data is transmitted in the sub-frame 210.
The PUSCH 240 is the only physical channel to which the UL shared (transport) Channel (UL-SCH) may be mapped. Therefore when a user has data to transmit and uses the transport channel UL-SCH, it must first gain access to the PUSCH 240 and to do so must inform the scheduler at the base station (eNB) 102 of this need.
A PUSCH transmission may carry an Uplink Shared Channel (UL-SCH) transport block which may include user plane data, control information (like MAC headers) and RRC signalling. FIG. 3 shows the construction of an UL-SCH MAC transport block 300, made up of a MAC header portion 310 and a MAC Payload portion 320 which itself may comprise MAC control elements 330, MAC service data units (SDUs) 340 and MAC padding bits 350
The shared PUSCH resource 240 is available (under control by the scheduler) for all commonly connected devices to use for the transmission of data. The UEs may indicate to the eNB their need to access the shared PUSCH resource 240 by one of several methods, including i) executing a random access procedure on a Physical Random Access Channel (PRACH), ii) transmitting a Buffer Status Report (BSR) on a PUSCH resource previously allocated via dynamic scheduling to the UE, or transmitting a Dedicated Scheduling Request (DSR) on a PUCCH resource 230. The eNB may use such indications when determining its allocation of the PUSCH resources 240.
The PUCCH resource 230 is semi-statically configured for reporting channel quality or channel status indicators (like CQI/PMI/RI), and for dedicated scheduling requests (DSR) to aid the eNB in its allocation of PUSCH resources 240. Portions of the PUCCH resource 230 are dynamically allocated for reporting ACK/NACK information. To achieve the dynamic allocation of PUCCH for ACK/NACK, the PUCCH resource used for a particular ACK/NACK transmission may be associated with the location of a corresponding DCI message on PDCCH.
The current LTE system is designed around the premise that connected mode is used only for UEs 101 with recent data activity. Therefore, a common assumption is that on entering RRC connected mode, a user will be semi-statically assigned (usually for the duration of the connected mode stay) dedicated SR resources on PUCCH for the purposes of informing the eNB 102 of the UEs 101 need to transmit data on PUSCH, subsequent to a new arrival of data and having previously had an empty transmission buffer.
Therefore, in the current “dedicated SR” approach for connected mode users, each user is assigned its own reserved time/frequency/code resource on uplink, on which the UE 101 may send a signal to indicate its need to access the PUSCH.
PUCCH Format 1
The signal commonly takes the form of PUCCH format 1, illustrated in FIG. 4, which is formed using a combination of frequency domain spreading 410 and time domain spreading 420 of a single-valued modulation symbol d(0) (d(0) is set to the value “1”). The single-valued modulation symbol d(0) is spread in both the time and frequency domains such that it occupies all of the REs within the sub-frame and resource block on those SC-FDMA symbols that are not used for RS 430.
The presence of PUCCH format 1 on the UEs dedicated PUCCH resource is sufficient to indicate to the eNB that a UE needs to access PUSCH. The absence of PUCCH format 1 is interpreted by the eNB as “no current need to access PUSCH”. Thus, PUCCH format 1 uses “on/off keying” to convey its information. Reference symbols 430 are inserted in the 3rd, 4th and 5th symbol locations of each slot (for a system employing normal cyclic prefix length).
There are also two additional variants of PUCCH format 1 that are defined within the standard, known as PUCCH format 1a and PUCCH format 1b. These have the same signal construction to that of PUCCH format 1 described above but allow for BPSK and QPSK modulation (respectively) of the symbol d(0). In this way, PUCCH format 1a may carry 1 bit of information (BPSK) and PUCCH format 1b may carry 2 bits of information (QPSK). These variants of PUCCH format 1 are used for transmission of Hybrid Automatic Repeat Request (HARQ) feedback, also known as ACK/NACK information.
PUCCH Formats 2 and 3
There are other PUCCH formats in the existing LTE specification, not used for DSR. PUCCH format 2 is used to carry Channel Quality Indicators (CQI), Precoding Matrix Indicators (PMI) and Rank Indicators (RI) for channel feedback purposes. PUCCH format 3 is also available for providing HARQ feedback (ACK/NACK) for PDSCH transmissions. Both of these formats can carry more data than PUCCH format 1.
DSR resources for different UEs may be multiplexed within the PUCCH resource in the time, frequency or code domains. Time multiplexing is achieved by assigning a sub-frame periodicity to each UE and assigning different sub-frame offsets to those UEs such that DSR transmissions occur at mutually exclusive times. Frequency multiplexing is achieved by assigning different resource blocks to different UEs such that DSR transmissions occur in mutually exclusive frequency regions. Code multiplexing is achieved by assigning different time and/or frequency-domain spreading sequences to different UEs such that DSR transmissions occur on different code resources which are separable by the eNB receiver. Combinations of time multiplexing, frequency multiplexing and code multiplexing may be used. A problem with the DSR scheme in PUCCH format 1 is that it does not scale well as the connected-mode user population is increased. As the connected mode population size increases, either the amount of total system UL resources that are needed (to reserve mutually exclusive PUCCH resources for each of the UEs) becomes excessive, or if the total DSR resources are constrained, the SR latency performance is then degraded (i.e. SR opportunities in time for a given UE necessarily become scarce due to the need to resort to large-scale time multiplexing of the users dedicated SR resources (e.g. a user's dedicated SR opportunity may arise only once every 40 or 80 ms or so).
The approach of relying on time-domain multiplexing to support a large connected mode population causes a problem in that access latency is then increased, since there is a high probability that the UE cannot send the SR immediately following the arrival of new data in the buffer. It is clear therefore that with the dedicated SR approach, a trade-off always exists between the amount of resources consumed or reserved, and the access latency. An example of this type of problem is shown in FIG. 5, where in this case, an SR opportunity 510 occurs only every 40 ms for a particular UE. If data arrives for UL transmission in-between the opportunities, the UEs need for UL transmission resources (on PUSCH) cannot be communicated to the eNB until the next opportunity—a time period of up to 40 ms.
Furthermore, when attempting to achieve low latency access, (requiring frequent DSR opportunities), the utilisation of those resources (i.e. occasions when a Scheduling Request (SR) is actually sent) reduces for a given offered SR load. For many common traffic profiles, a UE may send SR relatively infrequently and DSR resources are likely to be heavily underutilised. It can often be the case that less than 1% (even less than 0.1%) of the DSR resources are actually used by the UE to send SR and this may detract from overall system efficiency. If these unutilised resources were not reserved for DSR, they could have been reassigned for other purposes, such as for the transmission of user data or control data on UL, hence system capacity could have been improved.
Hence, it would be preferable to enable a scheduling request approach (intended primarily for connected mode users but not limited to such) which has the following attributes:                low access latency        use of an orthogonal multiple access scheme        efficient use of UL resources and high resource utilisationRACH        
One known alternative to the dedicated SR approach is the use of the existing random access procedure to inform the eNB of a UE's need for uplink resources. This is a multi-step procedure, as illustrated in FIG. 6, and is designed to transmit minimal information during the initial contention phase, step 610. In order to minimise the information sent, step 610 does not include the transmission of a user ID. In step 620, the eNB 102 responds with an uplink grant of resources for each detected preamble from step 610. Access contention may remain during message 3, step 630 and the contention resolution message of step 640 is used to resolve the contention between any users who selected the same initial preamble during step 610. The contention resolution message 640 may not be sent if the message 3 630 that is successfully decoded by the eNB was from a connected mode UE. In this case, the presence of the UL grant message 650 (addressed to the UEs C-RNTI) is sufficient to resolve contention and allow uplink data transmission 660.
Whilst the RACH procedure is relatively efficient, it does involve multiple steps and this can increase the access latency. Under some configurations, it also relies on a non-orthogonal multiple access scheme (which offers reduced capacity compared to the orthogonal PUCCH and PUSCH multiple access schemes). Furthermore, a portion of the resources on PRACH are expended in providing time guard regions and frequency guard regions to avoid interference from RACH users into other time/frequency UL resource regions of the system (such as PUCCH or PUSCH). Hence, again, resource usage efficiency for PRACH can be sub-optimal.
Furthermore, LTE PRACH occupies a wide bandwidth in the frequency domain (6 Physical Resource Blocks—PRBs). Thus, the provision of frequent RACH opportunities in time (to provide lower latency access) can then occupy a large proportion of the over uplink resource space. Therefore, the existing RACH procedure is designed more for initial access purposes and is not optimised for low-latency connected-mode scheduling request purposes.